Data from: The motion and tilts of subsurface floats due to surface waves
Data files
Mar 20, 2026 version files 220.08 MB
-
DataArchive.zip
220.07 MB
-
README.md
5.37 KB
Abstract
Subsurface and nearly neutrally-buoyant floats can be stable, well-behaved platforms for measuring ocean dynamics in the near-surface wave zone. In this work, we measure and model the wave-induced tilt of one such platform with the goal of understanding how to best optimize velocity measurements in the upper ocean from floats. We use data from Lagrangian floats built at APL/UW (Applied Physics Laboratory) and carrying a Nortek Signature1000 acoustic Doppler current profiler with an AHRS (Attitude and Heading Reference System). We analyze carefully chosen data segments where the wave-induced tilts are measurable yet small, and the float does not rotate significantly. We adopt a two-dimensional, depth-downwave coordinate system and assume small tilts and linear surface wave dynamics. By combining the constraints due to geometry, a wave-following float, and wave dynamics, we link measurements of both linear and angular acceleration to measure the tilts to an accuracy of a few tenths of a degree and simultaneously show that the data is consistent with the analysis assumptions. This analysis is further vetted by swinging the AHRS on pendulums in the laboratory. The same tests disturbingly indicate that the AHRS attitude readings can have large errors for tilts larger than a few degrees. The wave-induced tilt of the floats is predicted to an accuracy of about 10\% from the wave properties by a 3-parameter linear semi-empirical model calibrated with field data. The waves force tilt through their horizontal acceleration and through their strain exerting torques on the float. These floats are a somewhat underdamped oscillator (Quality Factor=3, resonance at 3 second period) and will exhibit a decaying oscillation of a few cycles when perturbed. This archive contains the data and programs necessary to reproduce the analysis described above.
Dataset DOI: 10.5061/dryad.70rxwdc96
Accurate measurements of the ocean near the surface often require a detailed understanding of the motion of the platform caused by waves. Here, we model the vertical and horizontal motion and tilts of a subsurface oceanographic float near the surface and tune the model parameters using field data aided by laboratory measurements. This was part of several large physical oceanography field campaigns to understand the behavior of the upper 100m of the ocean on horizontal scales of 100m to 100 km, with an emphasis on the smaller scales.
File: DataArchive.zip
Description: Archive in a compressed zip file
Within this archive, there are two directories:
Directory: Laboratory Data and Analysis
Fullfiles/* contain the full .mat files from which the test runs were extracted. The variables are:
- Mtime- Matlab time in ‘datenum’ format as days since '00-Jan-0000’. January 1 1970 has a value of 719529.
- Temperature- Temperature measured in degrees C. Floating point number.
- Pressure - Pressure measured in decibars. Floating point number.
- Heading, Pitch, Roll- orientation of the AHRS as computed by its internal algorithm. Their definition is shown in the supplementary materials of the paper. Floating point number.
- Magnetometer- output from 3 axis magnetometer as a 3xN matrix of 16 bit integers.
- Accelerometer- output from 3 axis accelerometer as a 3xN matrix of single precision floating point numbers in units of g, the acceleration of gravity.
- AHRSGyro - output from 3 axis rotation rate sensor as a 3xN matrix of single precision floating point numbers.
The conversions from the measurement of the magnetometer, accelerometer, and AHRSGyro to the analysis units and sign conventions are shown in program ‘MakeTiltspecPendulumS.m’ included in this dataset.
CalibrationRuns/ contains a subset of the above files .mat files for each of the actual data runs used in the analysis. Only data used in the analyses are included. The variables are:
- Mtime - same as BB.Mtime above, but only for the time of the data run.
- Accelerometer- same as BB.Mtime above
- Magnetometer- same as BB.Magnetometer above
- AHRSGyro- same as BB.AHRSGyro above
- Heading, Pitch, Roll - same as above
- CalibrationSeconds - time intervals chosen
- AngleToSwingPlaneDeg- rotate AHRS this amount to align with Pendulum swing angle, Degrees
- SigmaSwingFrequencyHz - frequency of the pendulum, Hz
- dtSamplingRateSec - sampling time seconds
- LPivotToAHRSm- Pendulum length, meters, measured from pivot to AHRS
PNG/ contains plots of the calibration runs (red) on top of the full files to provide context for the measurements.
Two MATLAB programs are included
- GetArchivePendulum.m is the script used to extract Calibration runs from the Fullfiles.
- MaketiltspecPendulumS.m is the script used to make auto and cross spectra for each calibration run.
Directory: Ocean Data and Analysis
There are two folders CALYPSO and LCDRI.
LCDRI/ contains data and programs for the Langmuir Cell Directed Research Initiative, on which most of the analysis is based. The programs will analyze the data and produce the major results of the paper.
5 folders starting with 100455-* each contain data for one float deployment.
There are 4-6 data files in each folder with names like
LCDRI_00_Average_corr.mat
LCDRI_01_Average_corr.mat
LCDRI_00_IBurst_corr.mat
LCDRI_01_IBurst_corr.mat
These contain the same variables as the Fullfiles/ but for the field data.
CDIP Data/ CDIP Data/* contains wave data from CDIP buoy station 229 in a fully documented netcdf file 229p1_d01.nc. The data is also archived at two sites
https://cdip.ucsd.edu/themes/cdip/?d2=p70:s:229
https://www.ncei.noaa.gov/thredds-ocean/catalog/sio/cdip/archive/229p1/catalog.html
Matlab code is included:
MakeTiltspec_Archive.m reads the ADCP data files and makes files of the auto and cross spectra.
FitModel_2025_April.m reads these and fits the model
TiltErrfun3.m is a function needed by FitModel_2025_April.m.
These will run and put files in directories SpecFiles/, SpecSamp/ and Fits/
CALYPSO/ contains data and programs for a single float deployment to contrast with those in LCDRI/*. The structure is the same
Data file- CALYPSO_2019_IBurst_corr.mat
Wave Data- WaveDirection.mat with variables
-
mtime - Matlab time
-
yd - yearday of 2019. January 1 is day 1.
-
th - angle of wave propagation. CCW from East in radians. Direction the waves are going.
Matlab code:
MakeTiltspec_CALYPSO19.m reads the ADCP data file and makes files of the auto and cross spectra.
FitModel_2024.m reads these and fits the model
TiltErrfun3.m is a function needed by FitModel_2024
These will run and put files in directories SpecFiles/, SpecSamp/ and Fits/
Human subjects data
The raw data contained no human identification information and does therefore not need to be de-indentified.
